International Journal of Computer Sciences and Engineering Open Access

Research Paper Vol.-6, Issue-9, Sept. 2018 E-ISSN: 2347-2693

An Improvement in Maximum Difference Method to Find Initial Basic

Feasible Solution for Transportation Problem
Lakhveer kaur1*, Madhuchanda Rakshit2, Sandeep Singh3
1,2
Department of Applied Science, Guru Kashi University, Talwandi Sabo, Bathinda, INDIA
3
Department of Mathematics, Akal University, Talwandi Sabo, Bathinda, INDIA

Available online at: www.ijcseonline.org

Accepted: 26/Sept/2018, Published: 30/Sept/2018

Abstract- It is very important to find initial solution of transportation problems to reach optimal solution. In this paper
Maximum Difference Method (MDM) is improved to get best initial solution of transportation problems. Our improved
method overcomes the limitations of MDM given by Smita Sood and keerti Jain. This modified approach most of times give
better solution than MDM specially in case of tie and very close to the optimal solution. Also sometimes gives optimal
solution.

AMS Subject Classification: 90C08, 90C05

Keywords: Transportation Problem, Optimal Solution, Initial Basic Feasible Solution, MDM

I. INTRODUCTION

The transportation problem (TP) is a special case of Linear Programming Problems in which commodities are transported from
several sources to different destinations in such a way that transportation cost should be minimum while satisfying both the
supply unit and the demand requirement. These transportation problems are usually solved by modified distribution method
(MODI) [3] or Stepping Stone Method [2]. To solve TP by using these methods, it is essential to proceed with initial basic
feasible solution (IBFS). Existing method to find IBFS are, NWCM, LCM and VAM [17], Improved versions of VAM [1, 4,
15, 19, 20], Total opportunity cost method [8], TOCM-VAM [12], MDM [21], JHM Method [5], ICVM [13] and more
methods to find IBFS are given by [6, 7, 9–11, 14, 16, 18]. In all these existing algorithms to find IBFS of transportation
problems one of the algorithm is developed by Smita Sood and Kirti Jain[21]which is named as Maximum Difference Method
(MDM). This method not always gives better solution because some of its limitations which are explained below: In
MDM,Penalties are calculated by the difference of maximum cost and next to maximum cost of transportation table in rows
and columns. if maximum and next maximum cost are same then penalties are taken to be zero which is seems to be
inappropriate because in many cases row or column having two maximum costs are equal but there is more difference between
maximum and next to the maximum (not same) cost. but allocation is given to another cell which may have comparatively
more cost. so in this case we do not get best feasible solution. Secondly, In MDM, If same maximum penalty in two or more
rows or columns is occur then top most row and extreme left column is selected for the allocation. but in the top most row and
extreme left corner if minimum cost is comparatively greater than the cost in row or column which is not on the top most row
and extreme left column then greater cost is selected for allocation and then transportation cost is not minimized in that case.

So main reason of all these limitations is no proper rule in MDM for presence of tie, which can be occur in selecting cost cells
for calculating penalties or for making allocations or in both cases. For removing all these limitations we have given
improvement in MDM.We have applied our modified algorithm over number of examples (specially those examples in which
tie occurs at different steps of solution processor).
In next sections modified algorithm is presented. Then numerical examples are provided and important observations are
summarized in last section. Optimal solution of Numerical examples is obtained by modified distribution method[3].

II. PROPOSED HEURISTIC

To modify MDM, Given steps are followed:

Step 1: Balance the transportation problem if it is unbalanced by adding dummy column with zero cost if total demand is less
than total supply and add dummy row if total supply is less than total demand.

© 2018, IJCSE All Rights Reserved 533

 International Journal of Computer Sciences and Engineering Vol.6(9), Sept 2018, E-ISSN: 2347-2693

Step 2: Identify the cells which have the maximum cost and next to the maximum (not same) cost of each row and calculate the
difference (penalty) between them and write it along the side of transportation table against corresponding row.
Step 3: Identify the cell which have the maximum cost and next to the maximum (not same) cost of each column and calculate
the difference (penalty) between them and write it below the transportation table against corresponding column.
Step 4: Identify maximum penalty, if it is below the table then make allocation in cell having minimum cost in the column
corresponding to the maximum penalty and if it is along the side of table then make allocation in the cell having minimum cost
in the row corresponding to maximum penalty.
Step 5: If maximum penalty in two or more rows or columns is equal,then make allocation in row or column having minimum
cost. if minimum cost is also same then choose the row or column in which maximum allocation can be given to the cell having
minimum cost.
Step 6: Adjust the supply and demand and cross out the satisfied row or column.if both row and column are satisfied
simultaneously then cross out only one of them and assign zero supply (or demand) in the remaining column (or row) . further
calculation is not required for the row or column which is crossed out.
Step 7: Calculate the penalties again for the remaining rows and columns by repeating the step2 and step 3 and make
allocations by repeating step 4 to step 6 until all the demand requirements and supply units are not satisfied.
Step 8: Calculate transportation cost by taking sum of product of the allocations and corresponding cells.

III. NUMERICAL EXAMPLES

Numerical examples are considered for the application of proposed modified method. So that modified method can be clearly
specified. Input data and results obtained by applying proposed approach and
MDM are given in table 1.

Table 1: Input Data for numerical examples and IBFS using Proposed Modified Approach and MDM

© 2018, IJCSE All Rights Reserved 534

 International Journal of Computer Sciences and Engineering Vol.6(9), Sept 2018, E-ISSN: 2347-2693

On solving above TPs by MDM, it is observed that there is no rule for selecting minimum cost cell (in case of tie) for making
allocation. For example, in example 2, there is a tie in selecting cell c31 and c41 for making allocation. If we first select cell
c31 for giving allocation according to MDM rule, i.e topmost row, then solution obtained as 114, which is not best solution.
But according to our proposed approach, we first select cell c41, to which maximum allocation is possible and obtained
solution is same as optimal solution.
IV. RESULT ANALYSIS

In this section, all the results obtained by solving TPs of Section 3 are interpreted and comparision is made between proposed
modified approach, MDM and VAM. Optimal solution is also obtained by MODI method. Input data is given in table 2.

V. Conclusion

In this paper, we have find out the limitations of MDM and improved it by removing its limitations so that best initial solution
for minimizing the cost of transportation problem should be obtained. This proposed modified approach gives better solution
than MDM when tie occurs in selecting cost cells and making allocations. Also most of times obtained solution is same as
optimal solution or very close to the optimal solution.

REFERENCES

[1] Balakrishnan N (1990) Mdified Vogels Approximation Method for the unbalanced transportation problem. Appl. Math. Lett. 3(2): 9- 11.
[2] Charnes A, Cooper WW (1954) The stepping-stone method for explaining linear programming calculations in transportation problems. Manag.
Sci. 1(1): 49- 69.
[3] Dantzig GB (1963) Linear Programming and Extensions. Princeton, NJ:Princeton University Press.
[4] Goyal SK (1984) Improving VAM for unbalanced transportation problems. J. Oper. Res. Soc. 35(12): 1113- 1114
[5] Juman ZAMS, Hoque MA (2015) An efficient heuristic to obtain a better initial feasible solution to the transportation problem. Appl Soft
Comput. 34: 813-826.
[6] Kasana HS, Kumar KD (2004) Introductory Operations Research:Theory and Applications. Springer, Heidelberg.
[7] Khan AR (2011) A re-solution of transportation problem:an algorithmic approach. Jahangirnagar University of journal of Science. 34(2): 49- 62.
[8] Kirca O, Satir A (1990) A heuristic for obtaining an initial solution for transportation problem. J. Oper. Res. Soc. 41(9); 865- 871.
[9] Korukoglu S, Balli S (2011) An improved Vogels Approximation method for the transportation problem. Mathematical and Computational
Applications. 16(2): 370- 381.
[10] Kulkarni SS (2012) On initial basic feasible solution for transportation problem-A new approach. J. Indian Acad. Math. 34(1): 19-25.
[11] Kulkarni SS, Datar HG (2010) On solution to modified unbalanced transportation problem. Bulletin of marathwada Mathematical Society. 11(2):
20- 26.
[12] Mathirajan M, Meenakshi B (2004) Experimental analysis of some variants of Vogels approximation method. Asia Pac J Oper Res. 21(4): 447-
462.
[13] Jude O, Ifeanyichukwu OB, Ihuoma IA, Akpos EP (2017) A new and efficient proposed approach to find initia l basic feasible solution of a
transportation problem. American Journal of Applied Mathematics and Statistics 5(2): 54-61
[14] Pargar F, Javadian N, Ganji AP (2009) A heuristic for obtaining an initial solution for the transportation problem with experimental analysis. The
6th International Industrial Engineering Conference, Sharif University of Technology, Tehran, Iran.
[15] Ramakrishnan CS (1988) An improvement to Goyals modified VAM for the unbalanced transportation problems. J. Oper. Res. Soc. 39(6): 609-
610.
[16] Rashid A, Ahmad SS, Uddin MS (2013) Development of a new heuristic for improvement of initial basic feasible solution of a balanced
transportation problem. Jahangirnagar University Journal of Mathematics and Mathematical Science. 28: 105-112.
[17] Reinfeld NV, Vogel WR (1958) Mathematical Programming. Englewood Cliffs. New Jersey:Prentice-Hall.
[18] Russell EJ (1969) Extension of Dantzigs algorithm to finding an initial near-optimal basis for transportation problem. Oper. Res. 17: 187-191.
[19] Shafaat A, Goyal SK (1988) Resolution of degeneracy in transportation problems. J. Oper. Res. Soc. 39(4): 411- 413.
[20] Shimshak DG, Kaslik JA, Barclay TD (1981) A modification of Vogels approximation through the use of heuristic. INFOR. 19: 259- 263.
[21] Sood S, Jain K. (2015) The Maximum Difference Method to find initial basic solution for Transportation Problem. Asian Journal of Managment
Science. 03(07): 08-11.

© 2018, IJCSE All Rights Reserved 535